In this application, we propose to investigate genomic heterogeneity within oral squamous cell carcinomas (OSCC) towards defining a cancer initiating cell (CIC) molecular signature that will advance our diagnostic and therapeutic armamentarium. Specifically, this work will address the NIDCR FOA on Functional Characterization of Oral Cancer Initiating Cells that emphasizes a critical need for the elucidation of specific OSCC CIC signature. CICs are proposed to contribute to treatment resistance and metastasis but many of their features are defined by phenotypes only and a thorough molecular understanding of what drives their aggressive behavior is lacking. We developed a mouse oral cancer (MOC) cell line model that has many conserved features with human OSCC. Overlapping features included CD44 expression and increased ERK1/2 activation in aggressive lines that paralleled human CIC populations. We translated these findings into a clinical trial where patients with OSCC receive trametinib, a MEK/ERK inhibitor, with the goal of targeting CIC populations. In addition, using exome sequencing of MOC lines, we found common driver pathways described in human OSCC. Expression microarray analysis then defined a MOC molecular signature of aggressiveness and metastasis that we call the Oral Cancer Metastasis Predictor (OCMP). This OCMP stratified human OSCC patients from three independent datasets comprising 321 patients into those with good outcomes and those with poor outcomes. Finally, we generated preliminary exome data from sorted human OSCC that in part paralleled findings in MOC lines. To test the hypothesis that the OCMP is a CIC specific imprint, we propose to use RNA capture sequencing to define the genomic heterogeneity within human OSCC. This approach will be the first to sort primary human OSCC into CIC and non-CIC populations and directly interrogate them with genomic approaches. These data will help refine the OCMP and likely identify other vulnerabilities within CIC populations. Similarly, we wil interrogate patient samples from the trametinib clinical trial, to assess whether MEK inhibition has an impact on the CIC signature within tumors. Finally, we will examine the functional contribution of Nkx2.3, a putative CIC specific transcription factor identified in the OCMP, to OSCC aggressiveness and metastasis. In particular, we will aim to identify Nkx2.3 targets that may be therapeutically relevant. Thus, using complementary approaches extending from a mouse model, to genomic approaches and including active CIC targeted clinical trial specimens, we propose to delineate the molecular underpinnings and signatures that may have diagnostic and therapeutic relevance in patients with OSCC.